© UNICEF Malawi/2012/Giacomo Pirozzi. Retrieved from: https://www.unicef.org/malawi/reports/unicef-malawi-country-programme-document
Predicting School Dropouts Using Machine Learning and Household Panel Data as an Early Warning System in Malawi
Dropping out of school is often not a sudden action but a culmination of various factors pushing children to leave school. For this reason, it is very important to pre-emptively identify students at risk and take preventive measures to improve school retention rates over time. Especially in modelling early warning systems, machine learning has recently proved to be a promising approach to identifying students at risk of dropout and informing preventive policy interventions for education staff and policymakers.¹ However, the current focus on using administrative data with ML applications also restricts the scope of these applications in data-scarce countries, where data collection and management systems are relatively more prone to financial and technical constraints.
A recent study released in the Journal of Computational Social Science (JCSS) by DA data scientist Hazal Colak Oz aims to address this gap in ML applications in low-income countries. The study focuses on developing predictive models for school dropout of primary school students in Malawi, the fourth poorest country in the world, with 70.3% of the population living below the extreme poverty line of US$1.90 per person per day.²
The results of this study show that predictive modelling using ML algorithms has the potential to identify students at risk of dropout at an early stage through accurate data and modelling, which can pave the way for integrated and scalable early warning systems for policymakers and education practitioners.
The findings of this study also highlight already-collected household data as important data sources to apply computational approaches in analysing education outcomes for children, especially in low-income and data-scarce countries. The methodological approach suggested in this study is also applicable in other countries where similar household panel datasets are available.
For further details on the technical and methodological aspects of this study, you can read the full article at this link.
Please email firstname.lastname@example.org if you would like to explore ways in which this methodology can be implemented in your country/policy context for an early warning system.
¹. Chung, J. Y., & Lee, S. (2018). Dropout early warning systems for high school students using machine learning. Children and Youth Services Review, 96, 346–353.
Kemper, L., Vorhoff, G., & Wigger, B. U. (2020). Predicting student dropout: A machine learning approach. European Journal of Higher Education, 10 (1), 28–47.
². World Health Organization (WHO). (2020). Population below international poverty line. Retrieved September 15, 2021 from: http:// uis. unesco. org/ sites/ defau lt/ files/ docum ents/ new- metho dologyshows-258- milli on- child ren- adole scents- and- youth- are- out- school. pd